Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. A method, comprising: providing a plurality of processing elements comprising one or more operators, the operators configured to process streaming data tuples by operation of one or more computer processors; establishing an operator graph of a plurality of operators, the operator graph defining at least one execution path in which a first operator of the plurality of operators is configured to receive data tuples from at least one upstream operator and transmit data tuples to at least one downstream operator; defining a breakpoint comprising a condition, the condition comprising a concurrent existence of a unique attribute value of data tuple in: (i) a first window of a first operator, and (ii) a second window of a second operator, wherein the first and second windows comprise a plurality of data tuples in the first and second operators, respectively; and upon detecting occurrence of the condition, triggering the breakpoint to halt processing by each of the plurality of operators in the operator graph.
A method for debugging streaming data processing applications involves defining operators that process data tuples. These operators are arranged in a graph where data flows from upstream to downstream operators. A breakpoint is set based on a condition: the simultaneous existence of a specific attribute value within a window of data tuples in one operator AND a window of data tuples in another operator. A window is simply a collection of data tuples at a given operator. When this condition is met, processing is halted across all operators in the graph, allowing for inspection and debugging.
2. The method of claim 1 , wherein triggering the breakpoint comprises: receiving a plurality of data tuples in the operator graph; monitoring the windows of the operators for occurrence of the condition; and upon detecting the occurrence of the condition, halting execution in each of the plurality of operators in the operator graph.
The method for debugging streaming data described in claim 1 halts processing by receiving data tuples, monitoring the windows of different operators to check if a unique attribute value exists concurrently in a first window of a first operator and a second window of a second operator, and when that condition is detected, stopping the execution of all operators in the operator graph.
3. The method of claim 1 , wherein the condition must occur within a specified amount of time, wherein the time is defined by a specified count of data tuples within the window of the respective operator.
The method for debugging streaming data described in claim 1 requires that the condition (the concurrent existence of a unique attribute value of a data tuple in a window of a first operator and in a window of a second operator) must be met within a specific timeframe. This timeframe is defined by the number of data tuples processed within the operator's window. So, the breakpoint is only triggered if the condition is met within, say, the last 100 tuples processed by each operator.
4. The method of claim 1 , wherein the first operator is in a first operator graph and the second operator is in a second operator graph, wherein triggering the breakpoint halts processing in the first operator graph and the second operator graph.
The method for debugging streaming data described in claim 1 allows for debugging across multiple independent operator graphs. Here, the first operator resides in a first operator graph, and the second operator resides in a second, distinct operator graph. When the breakpoint condition (the concurrent existence of a unique attribute value of a data tuple in a window of a first operator and in a window of a second operator) is met, processing halts in BOTH operator graphs, allowing coordinated debugging of interdependent data streams.
5. The method of claim 1 , further comprising imposing a window on an operator not configured to maintain a window of tuples.
The method for debugging streaming data described in claim 1 extends to debugging operators that don't inherently maintain a window of tuples. A window (a collection of data tuples) is artificially imposed on such an operator. This allows the debugging breakpoint (which relies on conditions within windows) to be applied even to operators that don't normally use windows.
6. The method of claim 5 , wherein imposing the window comprises utilizing a history of data tuples processed by the operator not configured to maintain the window of tuples.
The method for debugging streaming data described in claim 5, which imposes a window on an operator not configured to maintain one, does so by utilizing the history of data tuples that the operator has processed. This means the artificially imposed window is created by storing and retrieving recent tuples that flowed through the operator, allowing the breakpoint condition (requiring windows) to be evaluated.
7. A method, comprising: providing a plurality of processing elements comprising one or more operators, the operators configured to process streaming data tuples by operation of one or more computer processors; establishing an operator graph of a plurality of operators, the operator graph defining at least one execution path in which a first operator of the plurality of operators is configured to receive data tuples from at least one upstream operator and transmit data tuples to at least one downstream operator; defining define a breakpoint comprising a condition, the condition specifying that a result of a function applied against an attribute value of a data tuple in a window of an operator exceeds a first specified attribute threshold, wherein the window comprises a plurality of data tuples in the operator; and upon detecting occurrence of the condition, triggering the breakpoint to halt processing by each of the plurality of operators in the operator graph.
A method for debugging streaming data processing applications involves defining operators that process data tuples in a graph. A breakpoint is defined based on a condition: the result of a function applied to an attribute value of a data tuple in an operator's window exceeding a specified threshold. The window contains multiple data tuples. If this condition occurs, processing is halted across all operators in the graph. For example, the condition might be "average latency in the last 100 tuples exceeds 50ms".
8. The method of claim 7 , wherein the condition must occur within a specified amount of time, wherein the time is defined by a specified count of data tuples within the window of the operator, wherein the condition further specifies: (i) that the attribute value of the data tuple in the window of the operator exceeds a second specified attribute threshold, and (ii) the concurrent existence of a first data tuple in: (i) a first window of a first operator, and (ii) a second window of a second operator.
The method for debugging streaming data described in claim 7, where a breakpoint is triggered if a function applied to an attribute value in a data tuple within an operator's window exceeds a threshold, is refined further. The condition must occur within a specified timeframe (defined by a tuple count within the window). The condition also specifies that (i) the attribute value exceeds a *second* specified threshold, AND (ii) a specific tuple exists concurrently within a window of a first operator AND a window of a second operator.
9. The method of claim 8 , wherein triggering the breakpoint comprises: receiving a plurality of data tuples in the operator graph; monitoring the windows of the operators for occurrence of the condition; and upon detecting the occurrence of the condition, halting execution in each of the plurality of operators in the operator graph.
The method for debugging streaming data described in claim 8 halts processing by receiving data tuples in the operator graph, monitoring the windows of the operators to check if the breakpoint condition is met (where a function applied to an attribute value exceeds a first threshold, the attribute value exceeds a second threshold, and a specific tuple exists in windows of two different operators), and upon detecting that condition, stopping the execution of each operator in the graph.
10. The method of claim 9 , further comprising: upon determining that a third operator of the plurality of operators is not configured to maintain a window of data tuples: defining a window of the third operator comprising a plurality of attributes, the plurality of attributes comprising: (i) a window type, an eviction policy, and a trigger policy for the third window; and imposing the defined window on the third operator.
The method for debugging streaming data described in claim 9 extends to operators that don't maintain windows of data tuples. When an operator lacks a window, a window is defined for it, including attributes such as a window type (e.g., sliding, tumbling), an eviction policy (e.g., FIFO, LRU), and a trigger policy (e.g., based on time, tuple count). This defined window is then imposed on the operator.
11. A method, comprising: providing a plurality of processing elements comprising one or more operators, the operators configured to process streaming data tuples by operation of one or more computer processors; establishing an operator graph of a plurality of operators, the operator graph defining at least one execution path in which a first operator of the plurality of operators is configured to receive data tuples from at least one upstream operator and transmit data tuples to at least one downstream operator; defining define a breakpoint comprising a condition, the condition specifying that a first data tuple in a window of an operator has an attribute value violating a specified attribute threshold, wherein the window comprises a plurality of data tuples in the operator; and upon detecting occurrence of the condition, triggering the breakpoint to halt processing by each of the plurality of operators in the operator graph.
A method for debugging streaming data processing applications involves defining operators that process data tuples in a graph. A breakpoint is defined based on a condition: a data tuple in an operator's window has an attribute value that violates a specified threshold. The window contains multiple data tuples. If this condition occurs, processing is halted across all operators in the graph. Example: the "error_code" attribute in a tuple has a value of "500".
12. The method of claim 11 , wherein triggering the breakpoint comprises: receiving a plurality of data tuples in the operator graph; monitoring the windows of the operators for occurrence of the condition; and upon detecting the occurrence of the condition, halting execution in each of the plurality of operators in the operator graph.
The method for debugging streaming data described in claim 11 halts processing by receiving data tuples in the operator graph, monitoring the windows of the operators to check if the breakpoint condition is met (a tuple in a window has an attribute value violating a threshold), and upon detecting that condition, stopping the execution of each operator in the graph.
13. The method of claim 11 , wherein the condition must occur within a specified amount of time, wherein the time is defined by a specified count of data tuples within the window of the respective operator.
The method for debugging streaming data described in claim 11, where the breakpoint is triggered if an attribute value in a tuple's window violates a threshold, requires that the condition must be met within a specified timeframe. This timeframe is defined by the number of data tuples processed within the operator's window.
14. The method of claim 11 , wherein the attribute threshold indicates that the attribute value of the first tuple in a first window of a first operator graph cannot equal an attribute value of a second data tuple in a second window of a second operator graph, wherein triggering the breakpoint halts processing in the first operator graph and the second operator graph.
The method for debugging streaming data described in claim 11, where the breakpoint is triggered if an attribute value in a tuple's window violates a threshold, specifies that the attribute threshold indicates that the attribute value of a tuple in a window of a first operator graph *cannot* equal the attribute value of another tuple in a window of a *second* operator graph. When this inequality violation is detected, processing halts in BOTH operator graphs.
15. The method of claim 11 , further comprising imposing a window on an operator not configured to maintain a window of tuples.
The method for debugging streaming data described in claim 11 extends to debugging operators that don't inherently maintain a window of tuples. A window (a collection of data tuples) is artificially imposed on such an operator. This allows the debugging breakpoint (which relies on conditions within windows) to be applied even to operators that don't normally use windows.
16. The method of claim 1 , wherein the first and second operators are configured to sequentially process the data tuples in the first and second windows, wherein the data tuples in the first and second windows are duplicate data tuples, the duplicate data tuples comprising each attribute value of the data tuples in the first and second windows being identical.
The method for debugging streaming data described in claim 1 considers a scenario where the first and second operators process data tuples sequentially. In this case, the data tuples within their respective windows are duplicates, meaning each attribute value of the tuples in both windows is identical. This implies the debugging condition is triggered if the same data is seen in two places.
Unknown
December 30, 2014
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.